SEARCH · Search NASA

Results for “Distributed Memory Programming Models”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 55 records · Page 3

All sources NASA NTRS DOE OSTI

DOE OSTI2018

Roofline scaling trajectories: A method for parallel application and architectural performance analysis

The end of Dennard scaling signaled a shift in HPC supercomputer architectures from systems built from single-core processor architectures to systems built from multicore and eventually manycore architectures. This transition substantially complicated performance optimization and analysis as new programming models were created, new scaling methodologies deployed, and on-chip contention became a bottleneck to performance. Existing distributed memory performance models like logP and logGP were unable to capture this contention. The Roofline model was created to address this contention and its interplay with locality. However, to date, the Roofline model has focused on full-node concurrency. In this paper, we extend the Roofline model to capture the effects of concurrency on data locality and on-chip contention. We demonstrate the value of this new technique by evaluating the NAS parallel benchmarks on both multicore and manycore architectures under both strong-And weak-scaling regimes. In order to quantify the interplay between programming model and locality, we evaluate scaling under both the OpenMP and flat MPI programming models.

Ibrahim, K↗

NASA NTRS1994

Performance Measurement, Visualization and Modeling of Parallel and Distributed Programs

This paper presents a methodology for debugging the performance of message-passing programs on both tightly coupled and loosely coupled distributed-memory machines. The AIMS (Automated Instrumentation and Monitoring System) toolkit, a suite of software tools for measurement and analysis of performance, is introduced and its application illustrated using several benchmark programs drawn from the field of computational fluid dynamics. AIMS includes (i) Xinstrument, a powerful source-code instrumentor, which supports both Fortran77 and C as well as a number of different message-passing libraries including Intel's NX Thinking Machines' CMMD, and PVM; (ii) Monitor, a library of timestamping and trace -collection routines that run on supercomputers (such as Intel's iPSC/860, Delta, and Paragon and Thinking Machines' CM5) as well as on networks of workstations (including Convex Cluster and SparcStations connected by a LAN); (iii) Visualization Kernel, a trace-animation facility that supports source-code clickback, simultaneous visualization of computation and communication patterns, as well as analysis of data movements; (iv) Statistics Kernel, an advanced profiling facility, that associates a variety of performance data with various syntactic components of a parallel program; (v) Index Kernel, a diagnostic tool that helps pinpoint performance bottlenecks through the use of abstract indices; (vi) Modeling Kernel, a facility for automated modeling of message-passing programs that supports both simulation -based and analytical approaches to performance prediction and scalability analysis; (vii) Intrusion Compensator, a utility for recovering true performance from observed performance by removing the overheads of monitoring and their effects on the communication pattern of the program; and (viii) Compatibility Tools, that convert AIMS-generated traces into formats used by other performance-visualization tools, such as ParaGraph, Pablo, and certain AVS/Explorer modules.

Yan, Jerry C.↗

NASA NTRS1994

Discrete sensitivity derivatives of the Navier-Stokes equations with a parallel Krylov solver

This paper solves an 'incremental' form of the sensitivity equations derived by differentiating the discretized thin-layer Navier Stokes equations with respect to certain design variables of interest. The equations are solved with a parallel, preconditioned Generalized Minimal RESidual (GMRES) solver on a distributed-memory architecture. The 'serial' sensitivity analysis code is parallelized by using the Single Program Multiple Data (SPMD) programming model, domain decomposition techniques, and message-passing tools. Sensitivity derivatives are computed for low and high Reynolds number flows over a NACA 1406 airfoil on a 32-processor Intel Hypercube, and found to be identical to those computed on a single-processor Cray Y-MP. It is estimated that the parallel sensitivity analysis code has to be run on 40-50 processors of the Intel Hypercube in order to match the single-processor processing time of a Cray Y-MP.

Ajmani, Kumud↗

DOE OSTI2024

Porting ATLAS Fast Calorimeter Simulation to GPUs with Performance Portable Programming Models

FastCaloSim is a parameterized simulation of the particle energy response and of the energy distribution in the ATLAS calorimeter. It is a relatively small and self-contained package with massive inherent parallelism and captures the essence of GPU offloading via important operations like data transfer, memory initialization, floating point operations, and reduction. It was identified by the High Energy Physics Center for Computational Excellence project as a good testbed for evaluating the performance and ease of portability of programming models. In this paper, we will discuss the results of our evaluation of the porting process to Kokkos, SYCL, Alpaka, OpenMP and std::par (nvc++), and compare performance on NVIDIA, AMD and Intel GPUs, as well as multicore CPUs.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS1998

Performance Modeling and Measurement of Parallelized Code for Distributed Shared Memory Multiprocessors

This paper presents a model to evaluate the performance and overhead of parallelizing sequential code using compiler directives for multiprocessing on distributed shared memory (DSM) systems. With increasing popularity of shared address space architectures, it is essential to understand their performance impact on programs that benefit from shared memory multiprocessing. We present a simple model to characterize the performance of programs that are parallelized using compiler directives for shared memory multiprocessing. We parallelized the sequential implementation of NAS benchmarks using native Fortran77 compiler directives for an Origin2000, which is a DSM system based on a cache-coherent Non Uniform Memory Access (ccNUMA) architecture. We report measurement based performance of these parallelized benchmarks from four perspectives: efficacy of parallelization process; scalability; parallelization overhead; and comparison with hand-parallelized and -optimized version of the same benchmarks. Our results indicate that sequential programs can conveniently be parallelized for DSM systems using compiler directives but realizing performance gains as predicted by the performance model depends primarily on minimizing architecture-specific data locality overhead.

Waheed, Abdul↗

NASA NTRS1994

Implementation of a fully-balanced periodic tridiagonal solver on a parallel distributed memory architecture

While parallel computers offer significant computational performance, it is generally necessary to evaluate several programming strategies. Two programming strategies for a fairly common problem - a periodic tridiagonal solver - are developed and evaluated. Simple model calculations as well as timing results are presented to evaluate the various strategies. The particular tridiagonal solver evaluated is used in many computational fluid dynamic simulation codes. The feature that makes this algorithm unique is that these simulation codes usually require simultaneous solutions for multiple right-hand-sides (RHS) of the system of equations. Each RHS solutions is independent and thus can be computed in parallel. Thus a Gaussian elimination type algorithm can be used in a parallel computation and the more complicated approaches such as cyclic reduction are not required. The two strategies are a transpose strategy and a distributed solver strategy. For the transpose strategy, the data is moved so that a subset of all the RHS problems is solved on each of the several processors. This usually requires significant data movement between processor memories across a network. The second strategy attempts to have the algorithm allow the data across processor boundaries in a chained manner. This usually requires significantly less data movement. An approach to accomplish this second strategy in a near-perfect load-balanced manner is developed. In addition, an algorithm will be shown to directly transform a sequential Gaussian elimination type algorithm into the parallel chained, load-balanced algorithm.

Eidson, T. M.↗

NASA NTRS1988

Electronic Neural Networks

Memory based on neural network models content-addressable and fault-tolerant. System includes electronic equivalent of synaptic network; particular, matrix of programmable binary switching elements over which data distributed. Switches programmed in parallel by outputs of serial-input/parallel-output shift registers. Input and output terminals of bank of high-gain nonlinear amplifiers connected in nonlinear-feedback configuration by switches and by memory-prompting shift registers.

Lambe, John↗

DOE OSTI2020

GronOR: Massively Parallel and GPU-Accelerated Non-Orthogonal Configuration Interaction for Large Molecular Systems

GronOR is a program package for non-orthogonal configuration interaction calculations for an electronic wave function built in terms of anti-symmetrized products of multi-configuration molecular fragment wave functions. The two-electron integrals that have to be processed may be expressed in terms of atomic orbitals or in terms of an orbital basis determined from the molecular orbitals of the fragments. The code has been specifically designed for execution on distributed memory massively parallel and Graphics Processing Unit (GPU)-accelerated computer architectures, using an MPI+OpenACC/OpenMP programming approach. The task-based execution model used in the implementation allows for linear scaling with the number of nodes on the largest pre-exascale architectures available, provides hardware fault resiliency, and enables effective execution on systems with distinct central processing unit-only and GPU-accelerated partitions. The code interfaces with existing multi-configuration electronic structure codes that provide optimized molecular fragment orbitals, configuration interaction coefficients, and the required integrals. Algorithm and implementation details, parallel and accelerated performance benchmarks, and an analysis of the sensitivity of the accuracy of results and computational performance to thresholds used in the calculations are presented.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

NASA NTRS2003

Support of Multidimensional Parallelism in the OpenMP Programming Model

OpenMP is the current standard for shared-memory programming. While providing ease of parallel programming, the OpenMP programming model also has limitations which often effect the scalability of applications. Examples for these limitations are work distribution and point-to-point synchronization among threads. We propose extensions to the OpenMP programming model which allow the user to easily distribute the work in multiple dimensions and synchronize the workflow among the threads. The proposed extensions include four new constructs and the associated runtime library. They do not require changes to the source code and can be implemented based on the existing OpenMP standard. We illustrate the concept in a prototype translator and test with benchmark codes and a cloud modeling code.

Jin, Hao-Qiang↗

NASA NTRS2004

What Multilevel Parallel Programs do when you are not Watching: A Performance Analysis Case Study Comparing MPI/OpenMP, MLP, and Nested OpenMP

With the current trend in parallel computer architectures towards clusters of shared memory symmetric multi-processors, parallel programming techniques have evolved that support parallelism beyond a single level. When comparing the performance of applications based on different programming paradigms, it is important to differentiate between the influence of the programming model itself and other factors, such as implementation specific behavior of the operating system (OS) or architectural issues. Rewriting-a large scientific application in order to employ a new programming paradigms is usually a time consuming and error prone task. Before embarking on such an endeavor it is important to determine that there is really a gain that would not be possible with the current implementation. A detailed performance analysis is crucial to clarify these issues. The multilevel programming paradigms considered in this study are hybrid MPI/OpenMP, MLP, and nested OpenMP. The hybrid MPI/OpenMP approach is based on using MPI [7] for the coarse grained parallelization and OpenMP [9] for fine grained loop level parallelism. The MPI programming paradigm assumes a private address space for each process. Data is transferred by explicitly exchanging messages via calls to the MPI library. This model was originally designed for distributed memory architectures but is also suitable for shared memory systems. The second paradigm under consideration is MLP which was developed by Taft. The approach is similar to MPi/OpenMP, using a mix of coarse grain process level parallelization and loop level OpenMP parallelization. As it is the case with MPI, a private address space is assumed for each process. The MLP approach was developed for ccNUMA architectures and explicitly takes advantage of the availability of shared memory. A shared memory arena which is accessible by all processes is required. Communication is done by reading from and writing to the shared memory.

Jost, Gabriele↗

DOE OSTI2019

A Parallel Graph Environment for Real-World Data Analytics Workflows

Economic competitiveness and national security depend increasingly on the insightful analysis of large data sets. The diversity of real-world data sources and analytic workflows impose challenging hardware and software requirements for parallel graph platforms. The irregular nature of graph methods is not supported well by the deep memory hierarchies of conventional distributed systems, requiring new processor and runtime system designs to tolerate memory and synchronization latencies. Moreover, the efficiency of relational table operations and matrix computations are not attainable when data is stored in common graph data structures. In this paper, we present HAGGLE, a high-performance, scalable data analytics platform. The platform’s hybrid data model supports a variety of distributed, thread-safe data structures, parallel programming constructs, and persistent and streaming data. An abstract runtime layer enables us to map the stack to conventional, distributed computer systems with accelerators. The runtime uses multithreading, active messages, and data aggregation to hide memory and synchronization latencies on large-scale systems.

Graph Analytics↗

DOE OSTI2019

Exploring MPI Communication Models for Graph Applications Using Graph Matching as a Case Study

Traditional implementations of parallel graph operations on distributed memory platforms are written using Message Passing Interface (MPI) point to point communication primitives such as Send-Recv (blocking and nonblocking). Apart from this classical model, the MPI community has over the years added other communication models; however, their suitability for handling the irregular traffic workloads typical of graph operations remain comparatively less explored. Our aim in this paper is to study these relatively underutilized communication models of MPI for graph applications. More specifically, we evaluate MPI’s one-sided programming, or Remote Memory Access (RMA), and nearest neighborhood collectives using a process graph topology. There are features in these newer models that are intended to better map to irregular communication patterns, as exemplified in graph algorithms. As a concrete application for our case study, we use distributed memory implementations of a weighted graph matching algorithm to investigate performances of MPI-3 RMA and neighborhood collective operations compared to nonblocking Send- Recv. A matching in a graph is a subset of edges such that no two matched edges are incident on the same vertex. A maximum weight matching is a matching of maximum weight computed as the sum of the weights of matched edges. Algorithms for optimal matching have quadratic complexity and are difficult to parallelize. More importantly, execution of graph matching is dominated by high volume of irregular memory accesses, making it an ideal candidate for studying the effects of various MPI communication models on graph applications at scale. Our neighborhood collectives and RMA implementations yields up to 6× speedup over traditional optimized nonblocking Send-Recv implementations, on thousands of cores of the NERSC Cori supercomputer. We believe the lessons learned from this case study can be adopted to benefit a wider range of graph application codes.

Ghosh, Sayan↗

NASA NTRS1993

Automatic array alignment in data-parallel programs

FORTRAN 90 and other data-parallel languages express parallelism in the form of operations on data aggregates such as arrays. Misalignment of the operands of an array operation can reduce program performance on a distributed-memory parallel machine by requiring nonlocal data accesses. Determining array alignments that reduce communication is therefore a key issue in compiling such languages. We present a framework for the automatic determination of array alignments in array-based, data-parallel languages. Our language model handles array sectioning, reductions, spreads, transpositions, and masked operations. We decompose alignment functions into three constituents: axis, stride, and offset. For each of these subproblems, we show how to solve the alignment problem for a basic block of code, possibly containing common subexpressions. Alignments are generated for all array objects in the code, both named program variables and intermediate results. We assign computation to processors by virtue of explicit alignment of all temporaries; the resulting work assignment is in general better than that provided by the 'owner-computes' rule. Finally, we present some ideas for dealing with control flow, replication, and dynamic alignments that depend on loop induction variables.

Chatterjee, Siddhartha↗

NASA NTRS2006

Implementing Access to Data Distributed on Many Processors

A reference architecture is defined for an object-oriented implementation of domains, arrays, and distributions written in the programming language Chapel. This technology primarily addresses domains that contain arrays that have regular index sets with the low-level implementation details being beyond the scope of this discussion. What is defined is a complete set of object-oriented operators that allows one to perform data distributions for domain arrays involving regular arithmetic index sets. What is unique is that these operators allow for the arbitrary regions of the arrays to be fragmented and distributed across multiple processors with a single point of access giving the programmer the illusion that all the elements are collocated on a single processor. Today's massively parallel High Productivity Computing Systems (HPCS) are characterized by a modular structure, with a large number of processing and memory units connected by a high-speed network. Locality of access as well as load balancing are primary concerns in these systems that are typically used for high-performance scientific computation. Data distributions address these issues by providing a range of methods for spreading large data sets across the components of a system. Over the past two decades, many languages, systems, tools, and libraries have been developed for the support of distributions. Since the performance of data parallel applications is directly influenced by the distribution strategy, users often resort to low-level programming models that allow fine-tuning of the distribution aspects affecting performance, but, at the same time, are tedious and error-prone. This technology presents a reusable design of a data-distribution framework for data parallel high-performance applications. Distributions are a means to express locality in systems composed of large numbers of processor and memory components connected by a network. Since distributions have a great effect on the performance of applications, it is important that the distribution strategy is flexible, so its behavior can change depending on the needs of the application. At the same time, high productivity concerns require that the user be shielded from error-prone, tedious details such as communication and synchronization.

James, Mark↗

NASA NTRS1997

NAS Parallel Benchmark. Results 11-96: Performance Comparison of HPF and MPI Based NAS Parallel Benchmarks

High Performance Fortran (HPF), the high-level language for parallel Fortran programming, is based on Fortran 90. HALF was defined by an informal standards committee known as the High Performance Fortran Forum (HPFF) in 1993, and modeled on TMC's CM Fortran language. Several HPF features have since been incorporated into the draft ANSI/ISO Fortran 95, the next formal revision of the Fortran standard. HPF allows users to write a single parallel program that can execute on a serial machine, a shared-memory parallel machine, or a distributed-memory parallel machine. HPF eliminates the complex, error-prone task of explicitly specifying how, where, and when to pass messages between processors on distributed-memory machines, or when to synchronize processors on shared-memory machines. HPF is designed in a way that allows the programmer to code an application at a high level, and then selectively optimize portions of the code by dropping into message-passing or calling tuned library routines as 'extrinsics'. Compilers supporting High Performance Fortran features first appeared in late 1994 and early 1995 from Applied Parallel Research (APR) Digital Equipment Corporation, and The Portland Group (PGI). IBM introduced an HPF compiler for the IBM RS/6000 SP/2 in April of 1996. Over the past two years, these implementations have shown steady improvement in terms of both features and performance. The performance of various hardware/ programming model (HPF and MPI (message passing interface)) combinations will be compared, based on latest NAS (NASA Advanced Supercomputing) Parallel Benchmark (NPB) results, thus providing a cross-machine and cross-model comparison. Specifically, HPF based NPB results will be compared with MPI based NPB results to provide perspective on performance currently obtainable using HPF versus MPI or versus hand-tuned implementations such as those supplied by the hardware vendors. In addition we would also present NPB (Version 1.0) performance results for the following systems: DEC Alpha Server 8400 5/440, Fujitsu VPP Series (VX, VPP300, and VPP700), HP/Convex Exemplar SPP2000, IBM RS/6000 SP P2SC node (120 MHz) NEC SX-4/32, SGI/CRAY T3E, SGI Origin2000.

Saini, Subash↗

NASA NTRS1991

Computing Temperatures In Optically Pumped Laser Rods

Computer program presents new model solving temperature-distribution problem for laser rods of finite length and calculates both radial and axial components of temperature distributions in these rods. Contains several self-checking schemes to prevent over-writing of memory blocks and to provide simple tracing of information in case of trouble. Written in Microsoft FORTRAN 77.

Farrukh, Usamah O.↗

NASA NTRS1999

Research on Spectroscopy, Opacity, and Atmospheres

To make my calculations more readily accessible I have set up a web site cfaku5.harvard.edu that can also be accessed by FTP. it has 5 9GB disks that hold all of my atomic and diatomic molecular data, my tables of distribution function opacities, my grids of model atmospheres, colors, fluxes, etc, my program that are ready for distribution, most of my recent papers. Atlases and computed spectra will be added as they are completed. New atomic and molecular calculations will be added as they are completed. I got my atomic programs that had been running on a Cray at the San Diego Supercomputer Center to run on my Vaxes and Alpha. I started with Ni and Co because there were new laboratory analyses that included isotopic and hyperfine splitting. Those calculations are described in the appended abstract for the 6th Atomic Spectroscopy and oscillator Strengths meeting in Victoria last summer. A surprising finding is that quadrupole transitions have been grossly in error because mixing with higher levels has not been included. I now have enough memory in my Alpha to treat 3000 x 3000 matrices. I now include all levels up through n=9 for Fe I and 11, the spectra for which the most information is available. I am finishing those calculations right now. After Fe I and Fe 11, all other spectra are "easy", and I will be in mass production. ATL;LS12, my opacity sampling program for computing models with arbitrary abundances, has been put on the web server. I wrote a new distribution function opacity program for workstations that replaces the one I used on the Cray at the San Diego Supercomputer Center. Each set of abundances would take 100 Cray hours costing $100,000. 1 ran 25 cases. Each of my opacity CDs contains three abundances. I have a new program -iinning on the Alpha that takes about a week. I am going to have to get a faster processor or I will have to dedicate a whole workstation just to opacities.

Kurucz, Robert L.↗

DOE OSTI2023

Reinforcement Learning for Load-balanced Parallel Particle Tracing

We explore an online reinforcement learning (RL) paradigm to dynamically optimize parallel particle tracing performance in distributed-memory systems. Our method combines three novel components: (1) a work donation algorithm, (2) a high-order workload estimation model, and (3) a communication cost model. First, we design an RL-based work donation algorithm. Our algorithm monitors workloads of processes and creates RL agents to donate data blocks and particles from high-workload processes to low-workload processes to minimize program execution time. The agents learn the donation strategy on the fly based on reward and cost functions designed to consider processes' workload changes and data transfer costs of donation actions. Second, we propose a workload estimation model, helping RL agents estimate the workload distribution of processes in future computations. Third, we design a communication cost model that considers both block and particle data exchange costs, helping RL agents make effective decisions with minimized communication costs. We demonstrate that our algorithm adapts to different flow behaviors in large-scale fluid dynamics, ocean, and weather simulation data. Our algorithm improves parallel particle tracing performance in terms of parallel efficiency, load balance, and costs of I/O and communication for evaluations with up to 16,384 processors.

Distributed and parallel particle tracing↗